The analysis of the historical development of the currently used diagnosis schizophrenia in DSM or ICD classification indicates that sticking to established terms and methodologies more and more develops to an obstacle in schizophrenia research. The aim of the currently used operational diagnostic approach was to reach reliability of diagnoses between different psychiatrists and researchers as high as possible. This should improve the communication between different research groups and better the comparison of study results. The aim of a high reliability has surely been reached, however, the expected striking progress in research and clinical praxis failed to materialize till today. Reliability, which means agreement, does not simultaneously means validity. Modern brain and genetic research requires more homogenous clinical syndromes or ‘diseases’ to examine them with scientifically based methodologies and technologies. The prerequisite for doing this, however, is going back to clinical and empirical realities and to integrate these realities into the development of a new generation of classification systems that are free of dogmatic thinking.